20140517

RDF model

<mRNA> sio:has-ordered-part <p1>, <p2>, ..., <pE> ;
       sio:has-first-part <p1> ;
       sio:has-last-part <pE> .

<p1> sio:refers-to <exon1> ;
     sio:has-value "1"^^xsd:integer ;
     sio:is-before <p2> .

<p2> sio:refers-to <exon2> ;
     sio:has-value "2"^^xsd:integer ;
     sio:is-before <p3> ;
     sio:is-after <p1> .

Note: Use numerical IDs of SIO in real RDF data.

Ensembl

• Draft version of core RDF model is designed/implemented.
• Writing export script for trial tomorrow.
• Ensembl RDF will not use Identifiers.org URIs.
• Will use http://rdf.ebi.ac.uk/resource/ensembl/id and will be linked to Id.org URIs.

UniProt

• Changed UniProt model to match with Ensembl data model for cross links.

Identifiers.org

Variety of URIs that Identifiers.org provides:

• http://identifiers.org/db/id -- canonical URI
• http://info.identifiers.org/db[/id] -- provides info in (miriam) registry
• http://idtype.identifiers.org/db -- expresses "ID" is an identifier of the "db"

Modifications will be made to Identifiers.org:

• Will provide informative RDF for http://identifiers.org/db and http://identifiers.org/db/id
• Will remove trailing / from http://idtype.identifiers.org/db/

Identifiers.org URIs in RDF data

• We will use "http://identifiers.org/db/id sio:is-part-of|dct:isPartOf|void:inDataSet http://identifiers.org/db" instead of rdf:type.
• dct:isPartOf -- widely used
• void:inDataSet -- range must be :RDFDataset
• Checked xrefs in INSDC records /xref="DB:ID" for missing, discontinued, non-existing, unknown databases (rs_id.json, getentry_id.json)
• compare with 2100 namespaces in Bio2RDF

GFVO

• finished SIO mappings
• finished all comments (descriptions on how to use; what classes/properties actually mean)
• isDefinedBy for all classes (whether in GFF3, GVF, GTF, or VCF specification)
• migrated GFF3/GVF examples to the new refactored GFVO
• updated web-site at [http://www.biointerchange.org/ontologies.html]
• ontology live at [http://www.biointerchange.org/gfvo]